A network or system of computers often has instances of services dispersed on machines throughout the network. It is frequently desirable for the administrator of the network or system to be able to administer a particular service instance. For example, an administrator may need to modify or configure a particular DNS server, SQL Server, or Exchange Server.
Traditionally, in order for an administrator to locate and modify such service instances, it has been necessary for the administrator to have prior knowledge of the location of the particular service instance (or access to a service-specific directory of some sort) and knowledge of the particular tool or tools usable to modify the service instance.
There currently exist in certain contexts schemes for publishing the identification and location of network services via a directory. For example, in a particular distributed computing environment (DCE) utilizing Remote Procedure Call (RPC), there may exist a mechanism for publishing RPC servers in a directory. Also, the Novell Directory Service (NDS) has infrastructure for publishing network services in a directory. However, current tools which take advantage of these directories still require that the administrator, having used the directory to locate the service of interest, connect to the computer hosting the service and locate and invoke a tool usable to manage the service.
It can be readily appreciated that existing methods of network management do not allow an administrator to interact in an apparently direct manner with a service of interest, but rather force the administrator to interact with the service via the host computer and service-specific administrative tools. Accordingly, a method and system of network service administration is needed whereby a network administrator may administer a service of interest without prior knowledge of the location of the machine hosting the service or of the identification and mode of usage of the service-specific tools usable to administer the service.